Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Prostaglandin H synthase, also known as cyclooxygenase (COX) catalyzes the first rate-limiting step in the synthesis of prostaglandins and thromboxanes from arachidonic acid (Dewitt, 1991; Xie et al., 1991). Two COX isoenzymes have been identified: COX-1, which is constitutively expressed in a wide variety of tissues and is responsible for the low prostaglandin synthesis required for cell homeostasis (Dewitt, 1991; Pilbeam et al., 1993; Crofford, 1997); and COX-2, which is a highly-inducible enzyme that is expressed in the course of inflammation or other cellular stresses and accounts for the important synthesis of prostanoids that occurs in several physiopathological situations such as endotoxemia, septic shock, and local inflammation of target tissues (Kujubu et al., 1991; Feng et al., 1995; Dewitt, 1997). In addition to inflammation, elevated COX-2 expression has been associated with cell growth regulation and carcinogenesis (Crofford, 1997). Indeed, COX-2 behaves as an immediate-early gene inducible by lipopolysaccharide (LPS), cytokines, growth factors, and the tumor promoter TPA (Kujubu et al., 1991; Ryseck et al., 1992; Herschman et al., 1995).

COX-2 is also regulated at the transcription level by a diverse group of rodent liver tumor promoters known as peroxisomal proliferators (PPs). PPs cause an increase in the size and number of hepatic peroxisomes and enhance the expression of enzymes involved in fatty acid catabolism via activation of PP-activated receptors (PPARs), members of the steroid receptor superfamily (Green, 1995; Lee et al., 1995). Three subtypes of PPARs have been identified (, /, and ). PPAR is expressed in liver, gut, kidney, and brown adipose tissue (Kliewer et al., 1994; Braissant et al., 1996), whereas PPAR is found predominantly in white adipose tissue (Isseman and Green, 1990). PPAR is transcriptionally activated by hypolipidaemic fibrates such as clofibrate, fatty acids, and other ligands (Keller et al., 1993; Yu et al., 1995). Transcriptional regulation by PPARs is achieved through PPAR-retinoid X receptor heterodimers (where retinoid X receptor is the receptor for 9-cis-retinoic acid) which bind to DNA motifs (PPAR response elements) in the promoters of their target genes (Keller et al., 1993).

Hepatocytes respond well both in vivo and in vitro to most of the stimuli that positively regulate COX-2 expression in other cells (Dewitt, 1991; Herschman et al., 1995; Feng et al., 1995; Williams and DuBois, 1996). However, adult hepatocytes failed to express COX-2 regardless of the stimuli used and only Kupffer cells and immortalized mouse liver cells retain the ability to induce COX-2 (Zhang et al., 1995; Ledwith et al., 1997; Nanji et al., 1997).

Results from our group demonstrated that fetal hepatocytes of 21 days of gestation which exhibit a phenotype distinct from the adult counterparts respond well to LPS and proinflammatory cytokines (Casado et al., 1997). In view of these results, and taking into account that PPs have growth regulatory activities that are independent of peroxisome proliferation (Chen et al., 1994), we have investigated the effect of LPS and PPs on COX-2 expression and prostaglandin E₂ (PGE₂) synthesis in these hepatocytes. Our data show that COX-2 is expressed in these cells. However, in the presence of COX inhibitors, the amount of COX-2 but not the corresponding mRNA levels increased severalfold. Because these COX-2-specific inhibitors do not inhibit irreversibly the enzyme, these data indicate that pharmacological treatment with these drugs should consider the increase of COX-2 protein produced under these conditions and that it is catalytically active upon removal of the inhibitor.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. Clofibrate, prostaglandins, and LPS from Escherichia coli were purchased from Sigma Chemical Co. (St. Louis, MO). 15-deoxy-Prostaglandin J₂ was purchased from Calbiochem (Band Soden, Germany). Antibodies were obtained from Santa Cruz Laboratories (Santa Cruz, CA). Tissue culture dishes were obtained from Falcon (Lincoln Park, NJ). Tissue culture media were purchased from Biowhittaker (Walkersville, MD). Cytokines were from Boehringer (Mannheim, Germany). COX inhibitors were purchased from Cayman Chemical Co., Inc. (Ann Arbor, MI). PPs were obtained from Biomol Research Laboratories (Plymouth Meeting, PA). Reagents for electrophoresis were obtained from Bio-Rad (Hercules, CA).

Isolation and Culture of Fetal Hepatocytes. Hepatocytes from 21-day-old fetuses were prepared from pregnant albino Wistar rats (300-350 g). Animals were cared for following the Institutional Animal Care Instructions. Animals were fed on a standard laboratory diet and sacrificed between 09:00 and 10:00 h. Gestational age was assessed by standard criteria and fetuses were delivered by Caesarean section (Martín-Sanz et al., 1989). A suspension of fetal or neonatal hepatocytes was prepared by a nonperfusion collagenase dispersion method that involved incubation (3 g/flask) of chopped fetal liver for 30 min at 37°C with 15 ml of Ca²⁺-free Krebs-bicarbonate buffer containing 0.5 mM EGTA, under continuous gassing with a carbogen mixture (O₂/CO₂, 19:1; Martín-Sanz et al., 1989). The cell suspension was centrifuged (35g for 2 min) and the cell pellet was resuspended and incubated for 60 min with this medium containing 2.5 mM CaCl₂ and 0.5 mg/ml collagenase A (Boehringer Mannheim). At the end of the incubation period, the cells were centrifuged at 50g for 5 min and the cell pellet was resuspended and progressively filtered through nylon membranes of 500-µm, 100-µm and 50-µm mesh. Cell viability was assessed by trypan blue exclusion and was always higher than 90%. The hepatocyte suspension was washed twice with sterile Dulbecco's modified Eagle's medium (DMEM) and then kept in this medium supplemented with 50 µg/ml gentamicin, penicillin G, and streptomycin, respectively. Fetal hepatocytes were plated at 2 to 3 × 10⁶ cells in 6-cm tissue culture dishes containing 2.5 ml of DMEM supplemented with 10% of heat-inactivated fetal calf serum (FCS). Four hours after seeding the cells, the medium was aspirated and the plates were washed twice with PBS to remove the nonadherent cells. The hepatocytes were maintained in 2 ml of phenol red-free DMEM supplemented with 2% of heat-inactivated FCS . The amount of fetal hepatocytes expressing -fetoprotein was evaluated by immunocytochemistry and was higher than 90%.

Preparation of Peritoneal Macrophages. Adult male rats were maintained free of pathogens and bred in our colony. Resident peritoneal macrophages were prepared following a previous protocol (Casado et al., 1997). Briefly, light ether-anesthetized rats (4-6 animals) were sacrificed by cervical dislocation and injected i.p. with 15 ml of sterile RPMI 1640 medium. The peritoneal fluid was carefully aspirated to avoid hemorrhage and kept at 4°C to prevent the adhesion of the macrophages to the plastic. After centrifugation at 200g for 10 min at 4°C, the cell pellet was washed twice with 45 ml of ice-cold PBS. Cells were seeded at 2 × 10⁶ (6-cm dishes) with RPMI 1640 medium supplemented with 10% of heat-inactivated FCS. After incubation for 1 h at 37°C in 5% CO₂, nonadherent cells were removed by extensive washing with PBS. Cells were maintained in 2 ml of RPMI 1640 medium containing 10% of heat-inactivated FCS.

Preparation of Microsomal Fractions. Cells were washed twice with ice-cold PBS and homogenized with 1 ml of ice-cold extraction buffer (100 mM Tris-HCl, pH 7.4; 2 mM EDTA, 10 µg/ml leupeptin, 20 µg/ml aprotinin, and 0.5 mM phenylmethylsulfonyl fluoride) followed by three cycles of 15 s of sonication at 4°C. The homogenates were centrifuged at 10,000g for 15 min at 4°C. The resulting supernatants were centrifuged at 105,000g for 1 h at 4°C and the microsomal pellets were resuspended in buffer (20 mM Tris-HCl, pH 7.4; 0.2 mM dithiothreitol, and 0.5% Nonidet P-40). An aliquot was removed for protein determination (Bio-Rad protein reagent). Microsomes were boiled in Laemmli sample buffer (Laemmli, 1970) and microsomal protein (20 µg) was loaded to a 10% SDS-polyacrylamide gel electrophoresis, followed by Western blotting analysis (see below).

RNA Extraction and Analysis. Total RNA (3-4 ×10⁶ cells) was extracted following the guanidinium thiocyanate method (Chomczynski and Sacchi, 1987). After electrophoresis in a 0.9% agarose gel containing 2% formaldehyde the RNA was transferred to a Nytran membrane (NY 13-N; Schleicher & Schuell, Dassel, Germany) with 10× standard saline citrate (SSC; 10× SSC is 1.5 M NaCl/0.3 M sodium citrate, pH 7.4). The membrane was prehybridized and the level of COX-2 mRNA was determined by the corresponding full length cDNA as probe (Fletcher et al., 1992), labeled with [-³²P]dCTP with the Rediprime labeling kit (Amersham, Bucks, UK). The membrane was washed with 0.1× SSC and 0.1% SDS at room temperature for 10 min and twice at 42°C for 30 min. Quantification of the radioactive emission was performed in a Fuji BAS1000 detector (Kanagawa, Japan), avoiding saturation of the bands, and followed by exposure to X-ray film (Hyperfilm; Amersham). Normalization of the blots for RNA lane charge was performed by the hybridization with a probe specific for the 18S ribosomal RNA inserted into a PBR322 plasmid and labeled by nick translation.

Western Blot Analysis. The amount of COX-2 was determined in enriched microsomal preparations or in total cell extracts in the case of peritoneal macrophages. After determining the protein content, samples were boiled in Laemmli sample buffer and equal amounts of microsomal or soluble cell extract protein were size-fractionated in a 10% acrylamide gel, transferred to a polyvinylidene difluoride membrane (Amersham), and after blocking with 5% nonfat dry milk incubated with anti-COX-2 (1:1000) antibody from Santa Cruz Laboratories.

Determination of Metabolites. PGE₂ levels were determined in the culture medium with a specific enzyme immunoassay system and following the indications of the manufacturer (Amersham).

Data Analysis. The number of experiments is indicated in the corresponding figure. Statistical differences (P < .05) between mean values were determined by one-way ANOVA followed by Student's t test.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Fetal Hepatocytes Express COX-2 in Response to LPS: Modulation by COX Inhibitors. Primary cultures of hepatocytes prepared from fetal liver (21 days of gestation) release PGE₂ into the medium, reflecting the presence of a constitutive COX activity in these cells. When fetal hepatocytes were incubated with LPS, an important expression of COX-2 was observed as evidenced by the increase in PGE₂ synthesis (Fig. 1A) and the detection of both the immunoreactive protein in microsomal extracts (Western blot analysis) and the corresponding mRNA (Northern blot analysis) (Fig. 1B). When control or LPS-treated cells were incubated with indomethacin, a general COX inhibitor, or with the COX-2-specific inhibitor NS398 (Salvemini et al., 1995; Smith et al., 1996), the release of PGE₂ behaved differently depending on the stimulation of the hepatocytes (Table 1). Indomethacin inhibited more than 95% of the basal and LPS-induced activities. However, NS398 failed to inhibit PGE₂ production in unstimulated cells but blocked the synthesis in LPS-activated hepatocytes.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   LPS increases PGE2 synthesis in cultured hepatocytes after expression of COX-2. Cells (2-3 × 106) were stimulated with 1 µg/ml LPS and the release of PGE2 to the culture medium was determined (A). Cultured cells were stopped at 6 h to determine the amount of COX-2 mRNA by Northern blot, or after 24 h to determine the COX-2 immunoreactivity by Western blot, with a specific anti-COX-2 antibody. Hybridization with a ribosomal 18S probe was done to ensure the lane charge in the Northern blot (B). Results show the mean ± S.E.M. of three experiments assayed per triplicate (A; *P < .005 versus the control) or a representative experiment out of four (B).

View larger version (41K): [in this window] [in a new window]   Fig. 2.   COX-2 levels are increased in hepatocytes stimulated with LPS and treated with enzyme inhibitors. Cultured hepatocytes were treated with LPS (1 µg/ml), indomethacin, NS398 (both at 50 µM), or combinations of these. After 6 h of culture, the amount of COX-2 mRNA was determined by Northern blot after normalization for the content of ribosomal 18S RNA. After 24 h of culture, equal amounts of protein were analyzed by Western blot to determine the content of COX-2. Results show the mean ± S.E.M. of four experiments. *P < .001 versus cells treated with LPS in the absence of inhibitors).

View larger version (29K): [in this window] [in a new window]   Fig. 3.   Time course of COX-2 mRNA and protein levels in cells incubated with NS398. Hepatocytes were treated with LPS (1 µg/ml) in the absence or presence of 50 µM NS398. At the indicated times, cell extracts were prepared to determine the COX-2 mRNA and protein levels. The amount of COX-2 mRNA was expressed after normalization with the 18S ribosomal RNA. Results show the mean ± S.E.M. of three experiments. *P < .005 versus the corresponding condition in the absence of NS398.

PPAR Ligands Induce COX-2 Expression and Antagonize with LPS in this Process. Hepatocytes express PPAR that binds to the corresponding PPAR response elements sequence in the promoter regions of several genes (Braissant et al., 1996; Dowell et al., 1997). It has been described that PPs induce COX-2 expression in immortalized mouse liver cells (Ledwith et al., 1997). We analyzed the effect of clofibrate, a hypolipidemic compound, on COX-2 expression in our model of fetal hepatocytes. As Fig 4 shows, clofibrate increased COX-2 protein (at 24 h) and mRNA (at 6 h) in a dose-dependent manner. However, clofibrate assayed at concentrations higher than 700 µM resulted toxic for these cells (more than 40% of lactate dehydrogenase released to the medium at 1 mM clofibrate). The expression of COX-2 protein by clofibrate was always lower than that elicited by LPS (58% with respect to LPS).

View larger version (39K): [in this window] [in a new window]   Fig. 4.   Clofibrate induces COX-2 expression in fetal hepatocytes. Cultured cells (2-3 × 106) were stimulated with the indicated concentrations of clofibrate and the amount of COX-2 mRNA and protein were determined after 6 and 24 h of incubation, respectively. Results show the mean ± S.E.M. of three experiments. *P < .05; **P < .01;***P < .001 versus the condition in the absence of clofibrate.

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Effect of peroxisomal proliferators on COX-2 protein and PGE2 synthesis. Cultured hepatocytes were incubated for 24 h with the peroxisomal proliferators clofibrate, Wy14643, or ETYA (all assayed at 0.5 mM), and in the absence or presence of NS398 (50 µM). The amount of protein (A) and the release of PGE2 (B) were determined. LPS (1 µg/ml) was included to compare the effects with respect to the PPs. Simultaneous treatment of hepatocytes with LPS (1 µg/ml) and clofibrate resulted in an antagonistic response in terms of COX-2 protein accumulation (C). Results show the mean ± S.E.M. of three experiments. *P < .05; **P < .01;***P < .001 versus the corresponding condition in the absence of NS398 (A), or clofibrate (C).

View larger version (17K): [in this window] [in a new window]   Fig. 6.   The accumulation of COX-2 in the presence of NS398 is functional. Cultured hepatocytes were incubated for 20 h with LPS (1 µg/ml) or clofibrate (0.5 mM) and in the absence or presence of 50 µM NS398. After extensive washing of the cells with culture medium to remove these additions, the medium was replaced and cultured for an additional 4-h period. The amount of PGE2 released was determined. Results show the mean ± S.E.M. of three experiments assayed per triplicate. *P < .01; **P < .001 versus the condition in the absence of NS398.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this work we have analyzed the ability of fetal hepatocytes to express COX-2 in response to a proinflammatory stimuli such as LPS and to several PPs, as well as the effect of COX inhibitors on both COX-2 mRNA and protein levels in activated cells. Interestingly, fetal hepatocytes release moderate amounts of PGE₂ through COX-1 as deduced by pharmacological and biochemical criteria. Upon activation with LPS or PPs these cells express important levels of COX-2. Taken together, these results confirm COX-1 and COX-2 as the main prostaglandin synthesizing enzyme in control and LPS-stimulated hepatocytes, respectively. Moreover, LPS is more efficient than PPs inducing COX-2, and among the latter, clofibrate is the most potent PP expressing COX-2. However, ETYA and, to a lesser extent, Wy14643 are good activators of PPAR, at least 5- an 3-fold more effective than clofibrate, respectively, when analyzed on a coactivator-dependent receptor ligand assay (Krey et al., 1997). Therefore, it is possible that in addition to PPAR activation, clofibrate exerts other effects that favor COX-2 expression in these cells. Previous reports suggested that the expression of COX-2 elicited by clofibrate and other PPs yielded a virtually inactive enzyme (Ledwith et al., 1997). However, our results clearly show that the enzyme is functional if the concentration of PP used does not affect cell viability. Also, it is worth mentioning the antagonism between LPS and PPs in terms of COX-2 induction, which suggests the involvement of distinct pathways in the activation process. Analysis of this antagonism might offer new possibilities to manipulate pharmacologically the transcriptional control of this enzyme.

The accumulation of COX-2 protein observed when fetal hepatocytes are treated with COX inhibitors was unexpected and because of the magnitude of the effect, this result could be of considerable pharmacological interest, especially when reversible COX inhibitors are used. Moreover, not only in fetal hepatocytes, but also in macrophages, the inhibitors increased COX-2 protein levels, although the effect was quantitatively less important. The inducibility of COX-2 in macrophages has been well characterized (Yamada et al., 1997) and in agreement with our data, a moderate increase in the protein, but without changes at the mRNA level was evidenced after COX-2 inhibition with indomethacin in the macrophage cell line J774 (Pang and Hoult, 1996).

The mechanism by which indomethacin and NS398 favor COX-2 accumulation was partially analyzed. The data available indicated that the inhibitors did not affect COX-2 transcription, but rather acted at a post-translational level, either through stabilizing effects on the protein structure, or through the decrease in the synthesis of prostaglandins that favor COX-2 degradation or both. Among the metabolites assayed, only incubation of fetal hepatocytes with 15-deoxy-PGJ₂ resulted in a low but significant decrease in the amount of COX-2 protein, suggesting that protein stabilization by the inhibitors was the main contributor to COX-2 accumulation. It should be mentioned that a destabilizing effect of prostaglandins (mainly by PGE₂) on COX-2 protein has been described in macrophages by other groups (Pang and Hoult, 1996).

In addition to the effects of COX inhibitors on protein levels, it has been described that high concentrations of indomethacin and other nonsteroidal anti-inflammatory drugs, although they completely inhibited prostaglandin synthesis, induced COX-2 expression exhibiting a similar time course and dose response as that elicited by PPs (Ledwith et al., 1997; Lehman et al., 1997). This is because these drugs are potent activators of both PPAR and PPAR, therefore mimicking most of the effects of the PPs. However, this is not the mechanism of action observed in our experimental system because neither inhibitor by itself induced COX-2, and the amount of COX-2 mRNA in activated cells was not increased by these drugs, indicating that at the concentrations used they act on the protein stability.

Regarding the effects of prostaglandins on liver function, it has been reported that exogenous PGE₂ stimulated hepatocyte growth factor production by several cells and that this is an important cytokine for hepatocyte growth (Bamba et al., 1998). Indeed, indomethacin and other COX inhibitors decreased the release of hepatocyte growth factor, and together with other data, this suggests that prostaglandins are important regulators of normal hepatocyte development (Bamba et al., 1998). Also, interleukin-6, an important cytokine involved in the acute phase response and in liver regeneration is elevated by PGE₂ (Hinson et al., 1996).

With respect to COX-2 inducibility in the hepatocyte, previous work indicated that these cells, at least from adult animals, are extremely resistant to induce COX-2 in response to a wide array of treatments (Johnston and Kroening, 1996). Only in simian virus 40-transformed hepatocytes or in immortalized liver cells from adult mice has it been possible to observe COX-2 expression in response to phorbol esters, PPs, and proinflammatory factors (Ledwith et al., 1997; Yu et al., 1998).

In conclusion, it is presumed that COX-2 inhibitors favor a conformational change in the protein that substantially alters its normal turnover and differences probably exist among distinct cell types. This is of pharmacological relevance because the accumulated enzyme is fully active once the concentration of the inhibitor decreases. In this regard, the design and use of irreversible inhibitors of COX-2 might benefit a better control on COX activity. The use of aspirin-like molecules such as o-(acetoxyphenyl)hept-2-ynyl sulfide, which is 60-fold more reactive than aspirin for COX-2 (Kalgutkar et al., 1998) and that inactivate irreversibly the enzyme are the likely candidates to avoid this increased COX-2 activity.